Applied Mechanics and Materials Vols. 602-605

Abstract: The computer vision technology is an important branch of computer science and artificial intelligence which is regarded as a non-destructive testing technique in the field of agriculture with a broad application prospects. This paper introduces the application of the computer vision technology in the agricultural products deterioration recognition, builds foundations for the accurate measurement of the agricultural products quality with computer visions, and establish the relationship between the feature information and quality of the agricultural products. Meanwhile, this paper combined the computer vision technology with infrared, microwave, NMR techniques to extract and test the visual information of the internal quality of the agricultural products.

Abstract: in the process of computer network communication, different signals subject to different communication protocols and with different signal characteristics. After invading, an entangled state which cannot be constraint formed between the signals. The characteristic differences in the polarization domain between invasion signal and communication signal is significantly reduced, and it is not easy to purify. To solve this problem, in this paper, an novel intrusion signal classification model of orthogonal polarization array is established. According to the signal loss characteristics changed by the invasion polarization, it can achieve the characteristics separation of desired signal and interference signal in the region of the polarization. the interference immunity of the separated signal is performed by a simulation analysis, it can proved that taking this vector signal processing approach not only extends the capabilities of invasion signal processing in multi-signal communications, and can significantly improve the performance of interference resistance of multi-signal network communication system.

Abstract: In fault detection process of large tanning machine, fault signal fluctuations are susceptibly caused by interference of external environment. The traditional methods are difficult to accurately classify fault detection of such random fluctuations, resulting in latter detection with low accuracy. To avoid these shortcomings, support vector algorithm based on least squares is proposed for fault detection of large tanning machine. Experimental results show that the algorithm can improve the accuracy of fault detection.

Abstract: This paper focused on the fire detecting problem. Traditionally, fire was detected based on the frame difference by subtracting the image pixel. When there existed background similar to its flame color and shape in the environment where the fire happened,, the result of frame difference subtracting is not obvious and the algorithm cannot detect the fire problem according to the result. This paper put forward a fire detecting method based on support vector basis algorithm. The experiment indicated that this kind of neural network model achieved precise fire detecting under the background similar to itself, efficiently decreased the detecting error rate, and obtained satisfactory results.

Abstract: Due to the objects in the embedded control procedure are difficult to obtain a variety of fault data and fault features, it’s necessary to establish simulation models in accordance with the operational mechanisms of the embedded equipment to simulate and diagnose the practical faults. This paper proposes a SVM integrated diagnostic method and further proposes the faults classification model with improved neural network. The faults diagnose performance is greatly improved by analyzing the types of the faults in different facets. For the embedded valve failure modes, the simulation results of the proposed method are compared with that of the previous mature independent element analysis method. The simulation results show that the fault diagnosis method in this paper can effectively improve the speed and accuracy of fault diagnosis for the embedded equipment.

Abstract: The large-scale software is consisted of the components which are quite different. The detection accuracy of the traditional faults detection methods for the large-scale component software is not satisfactory. This paper proposes a large-scale software faults detection methods based on improved neural network combining the features of the large-scale software by computing the stable probability and building the neural network faults detection models. The proposed method can analyze the serial faults of the large-scale software to determine the positions of the faults. The experiment and simulation results show that the improved method for large-scale software fault detection can greatly improve the accuracy.

Abstract: Mountainous environment makes the wireless sensor network (WSN) data collection applications remain a challenging domain, as the detection region may present a three-dimensional structure and the radio propagation characteristics are still looking forward to further research. To better adapt to the ecological data acquisition needs in mountainous orchard, A WSN clustering data acquisition system is designed and implemented. It uses the received signal strength indicator (RSSI) to evaluate radio propagation performance and characterize the communication quality of the link. In the selection of the cluster heads and the next routing hops, this system takes RSSI, node’s residual energy and other influencing factors into account and use the multiple-attribute comprehensive evaluation model to weigh them comprehensively. Simulation results indicate that such a design can give objective and reasonable evaluations and judgments of the candidate nodes. Analyses verify the effectiveness and reasonability of the proposed model.

Abstract: Aiming at the shortages of traditional method for power transformer fault diagnosis, the ensemble idea and incremental learning idea are used for better performance. The SVM is selected to establish the fault diagnosis models as sub learning machines. And then, the Learn++ algorithm is used to aggregate the sub learning machines. The new with new method will ensure the accuracy of fault diagnosis, and will update online. The experiments demonstrate that the performance of power transformer fault diagnosis system based on Learn++ is the best.

Abstract: Now, the measure of network communications mainly adopts the measure of software focus which is based on the servers of telecom dealer. With the rapid increase of cutwork telecom users, the style we mentioned above ties up more and more hardware resource and network bandwidth, to telecom customers, this style measure is low at diaphaneity and hard to ensure the legitimacy rights. The “Measurement Meter for the Network Traffic” we research in this paper adapts the traffic measure and be used between the user’s port and the telecom network. It also compensated the measure’s weaknesses of unifying background measure by telecom dealer.

Abstract: Camshift, namely "Continuously Adaptive Mean-Shift" algorithm, is an adaptive tracking algorithm. This algorithm is based on the color information to track the moving target in image sequence. In the simple background, this algorithm achieved a steady and current tracking effect. But in dynamic scene, the global motion caused by the camera, the background of the light and occlusion will affect the accuracy, or even lose the tracking of the target. In order to solve the above problem, this paper adjust the H component in HSV color space, as well use weighted color histogram to improve the Camshift algorithm, then combined with Kalman filter to track the target in the image sequence. The experimental result shows that this approach can track object stability and correctly in dynamic scene.